U.S. patent number 10,789,296 [Application Number 16/031,811] was granted by the patent office on 2020-09-29 for detection of missing entities in a graph schema.
This patent grant is currently assigned to International Business Machines Corporation. The grantee listed for this patent is International Business Machines Corporation. Invention is credited to Charles E. Beller, Shaila Pervin, Cesar Augusto Rodriguez Bravo, Craig M. Trim.
![](/patent/grant/10789296/US10789296-20200929-D00000.png)
![](/patent/grant/10789296/US10789296-20200929-D00001.png)
![](/patent/grant/10789296/US10789296-20200929-D00002.png)
![](/patent/grant/10789296/US10789296-20200929-D00003.png)
United States Patent |
10,789,296 |
Trim , et al. |
September 29, 2020 |
Detection of missing entities in a graph schema
Abstract
Detecting missing entities in a graph schema is provided.
Unlabeled and unstructured data of a particular domain are divided
into a plurality of Voronoi cells using a clustering algorithm that
is initiated with cluster centroids proportional to a number of
entity types corresponding to the particular domain existing in the
graph schema. One additional cluster more than the number of entity
types corresponding to the particular domain existing in the graph
schema is initialized using a cluster initializing formula of the
clustering algorithm. It is determined whether the one additional
cluster is populated. In response to determining that the one
additional cluster is populated, an entity type is determined to be
missing from the number of entity types existing in the graph
schema. The missing entity type is added to the graph schema.
Inventors: |
Trim; Craig M. (Ventura,
CA), Beller; Charles E. (Baltimore, MD), Pervin;
Shaila (Docklands, AU), Rodriguez Bravo; Cesar
Augusto (Alajuela, CR) |
Applicant: |
Name |
City |
State |
Country |
Type |
International Business Machines Corporation |
Armonk |
NY |
US |
|
|
Assignee: |
International Business Machines
Corporation (Armonk, NY)
|
Family
ID: |
1000005083297 |
Appl.
No.: |
16/031,811 |
Filed: |
July 10, 2018 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20200019647 A1 |
Jan 16, 2020 |
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06K
9/6223 (20130101); G06N 20/00 (20190101); G06F
16/9024 (20190101); G06F 16/212 (20190101) |
Current International
Class: |
G06F
15/16 (20060101); G06N 20/00 (20190101); G06F
16/901 (20190101); G06K 9/62 (20060101); G06F
16/21 (20190101) |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
Jain, A. Data clustering: 50 years beyond K-means. Pattern
Recognition Letters 31 (2010), pp. 651-666. (Year: 2010). cited by
examiner .
Jiang, et al. Two-phase clustering process for outliers detection.
Pattern Recognition Letters 22 (2001), pp. 691-700. (Year: 2001).
cited by examiner .
Neelakantan et al., "Inferring Missing Entity Type Instances for
Knowledge Base Completion: New Dataset and Methods," arXiv preprint
arXiv:1504.06658 (2015). Retrieved from Internet using:
https://arxiv.org/pdf/1504.06658. cited by applicant .
Paulheim, "Knowledge Graph Refinement: A Survey of Approaches and
Evaluation Methods," Semantic Web 0, No. 3 (2017): pp. 489-508.
cited by applicant .
Wan et al., "An Approach of Entity Alignment Based on Semantic
Features," 2017 4th International Conference on Information,
Cybernetics and Computational Social Systems (ICCSS), 2017, pp.
170-174. cited by applicant .
Moon et al., "Learning Entity Type Embeddings for Knowledge Graph
Completion," CIKM '17 Proceedings of the 2017 ACM Conference on
Information and Knowledge Management, Nov. 2017, Singapore, pp.
2215-2218. cited by applicant .
Lin et al., "Learning Relational Bayesian Classifiers from RDF
Data," Proceedings of the 10th International Conference on the
Semantic Web, Oct. 23-27, 2011, Bonn, Germany, 16 pages. cited by
applicant.
|
Primary Examiner: Gofman; Alex
Assistant Examiner: Qian; Shelly X
Attorney, Agent or Firm: Yee & Associates, P.C.
Claims
What is claimed is:
1. A computer-implemented method for detecting missing entities in
a graph schema, the computer-implemented method comprising:
dividing unlabeled and unstructured data of a particular domain
into a plurality of Voronoi cells using a clustering algorithm that
is initiated with cluster centroids proportional to a number of
entities corresponding to the particular domain existing in the
graph schema; initializing one additional cluster more than the
number of entities corresponding to the particular domain existing
in the graph schema using a cluster initializing formula of the
clustering algorithm; populating clusters corresponding to entities
of the particular domain based on dividing the unlabeled and
unstructured data into the plurality of Voronoi cells; determining
whether the one additional cluster is populated; responsive to
determining that the one additional cluster is populated,
determining that an entity is missing from the number of entities
existing in the graph schema and adding the missing entity to the
graph schema; and responsive to determining that the one additional
cluster is not populated, determining that all entities of the
particular domain exist in the graph schema.
2. The computer-implemented method of claim 1 further comprising:
populating a number of clusters corresponding to the incremented
number of entities; determining whether a cluster in the number of
clusters is not populated; responsive to determining that the
number of clusters is populated, determining that another entity is
missing from the graph schema; and responsive to determining that a
cluster in the number of clusters is not populated, determining
that all entities of the particular domain exist in the graph
schema.
3. The computer-implemented method of claim 1 further comprising:
sending a notification regarding the missing entity in the graph
schema; and outputting the graph schema having all entities
corresponding to the particular domain modeled in the graph
schema.
4. The computer-implemented method of claim 3, wherein the graph
schema having all of the entities corresponding to the particular
domain modeled in the graph schema is outputted to a graph database
server hosting a graph database that complies with the graph schema
to increase query performance and efficiency of the graph database
server.
5. The computer-implemented method of claim 1 further comprising:
ingesting the unlabeled and unstructured data corresponding to the
particular domain from a plurality of data sources via a network;
and analyzing the unlabeled and unstructured data corresponding to
the particular domain using machine learning.
6. The computer-implemented method of claim 1, wherein the
clustering algorithm is a k-means clustering algorithm.
7. The computer-implemented method of claim 1, wherein the
clustering initializing formula defines that a number of clustering
centroids is proportional to a number of root entities plus
one.
8. A computer system for detecting missing entities in a graph
schema, the computer system comprising: a bus system; a storage
device connected to the bus system, wherein the storage device
stores program instructions; and a processor connected to the bus
system, wherein the processor executes the program instructions to:
divide unlabeled and unstructured data of a particular domain into
a plurality of Voronoi cells using a clustering algorithm that is
initiated with cluster centroids proportional to a number of
entities corresponding to the particular domain existing in the
graph schema; initialize one additional cluster more than the
number of entities corresponding to the particular domain existing
in the graph schema using a cluster initializing formula of the
clustering algorithm; populate clusters corresponding to entities
of the particular domain based on dividing the unlabeled and
unstructured data into the plurality of Voronoi cells; determine
whether the one additional cluster is populated; determine that an
entity is missing from the number of entities existing in the graph
schema and add the missing entity type to the graph schema in
response to determining that the one additional cluster is
populated; and determine that all entities of the particular domain
exist in the graph schema in response to determining that the one
additional cluster is not populated.
9. The computer system of claim 8, wherein the processor further
executes the program instructions to: populate a number of clusters
corresponding to the incremented number of entities; determine
whether a cluster in the number of clusters is not populated;
determine that another entity is missing from the graph schema in
response to determining that the number of clusters is populated;
and determine that all entities of the particular domain exist in
the graph schema in response to determining that a cluster in the
number of clusters is not populated.
10. The computer system of claim 8, wherein the processor further
executes the program instructions to: send a notification regarding
the missing entity in the graph schema; and output the graph schema
having all entities corresponding to the particular domain modeled
in the graph schema.
11. The computer system of claim 10, wherein the graph schema
having all of the entities corresponding to the particular domain
modeled in the graph schema is outputted to a graph database server
hosting a graph database that complies with the graph schema to
increase query performance and efficiency of the graph database
server.
12. The computer system of claim 8, wherein the processor further
executes the program instructions to: ingest the unlabeled and
unstructured data corresponding to the particular domain from a
plurality of data sources via a network; and analyze the unlabeled
and unstructured data corresponding to the particular domain using
machine learning.
13. A computer program product for detecting missing entities in a
graph schema, the computer program product comprising a computer
readable storage medium having program instructions embodied
therewith, the program instructions executable by a computer to
cause the computer to perform a method comprising: dividing
unlabeled and unstructured data of a particular domain into a
plurality of Voronoi cells using a clustering algorithm that is
initiated with cluster centroids proportional to a number of
entities corresponding to the particular domain existing in the
graph schema; initializing one additional cluster more than the
number of entities corresponding to the particular domain existing
in the graph schema using a cluster initializing formula of the
clustering algorithm; populating clusters corresponding to entities
of the particular domain based on dividing the unlabeled and
unstructured data into the plurality of Voronoi cells; determining
whether the one additional cluster is populated; responsive to
determining that the one additional cluster is populated,
determining that an entity is missing from the number of entities
existing in the graph schema and sending a notification regarding
the missing entity type; and responsive to determining that the one
additional cluster is not populated, determining that all entities
of the particular domain exist in the graph schema.
14. The computer program product of claim 13 further comprising:
populating a number of clusters corresponding to the incremented
number of entities; determining whether a cluster in the number of
clusters is not populated; responsive to determining that the
number of clusters is populated, determining that another entity is
missing from the graph schema; and responsive to determining that a
cluster in the number of clusters is not populated, determining
that all entities of the particular domain exist in the graph
schema.
15. The computer program product of claim 13 further comprising:
adding the missing entity to the graph schema; and outputting the
graph schema having all entities corresponding to the particular
domain modeled in the graph schema.
16. The computer program product of claim 15, wherein the graph
schema having all of the entities corresponding to the particular
domain modeled in the graph schema is outputted to a graph database
server hosting a graph database that complies with the graph schema
to increase query performance and efficiency of the graph database
server.
17. The computer program product of claim 13 further comprising:
ingesting the unlabeled and unstructured data corresponding to the
particular domain from a plurality of data sources via a network;
and analyzing the unlabeled and unstructured data corresponding to
the particular domain using machine learning.
18. The computer program product of claim 13, wherein the
clustering algorithm is a k-means clustering algorithm.
19. The computer program product of claim 13, wherein the
clustering initializing formula defines that a number of clustering
centroids is proportional to a number of root entities plus one.
Description
BACKGROUND
1. Field
The disclosure relates generally to graph databases and more
specifically to detecting missing entities in a graph schema of a
graph database and automatically adding the detected missing
entities to the graph schema.
2. Description of the Related Art
Structured data refer to data that are of a determined length or
structure and that reside in a fixed field or record. Credit card
numbers, social security numbers, and telephone numbers are but a
few examples of structured data. Structured data generally resides
in a relational database in a table format.
Unstructured data are information that either does not have a
pre-defined data model or is not organized in a pre-defined manner.
Unstructured data are typically in the form of textual information,
but may contain dates and numbers as well. This results in
irregularities and ambiguities that make unstructured data
difficult to understand using traditional programs as compared to
data stored in rows and columns in a traditional relational
database. Examples of unstructured data may include books,
journals, technical specifications, training manuals, product
catalogs, web pages, blogs, social media posts, documents,
metadata, records, audio files, video files, images, graphics,
emails, text messages, and the like. Increasingly, unstructured
data is becoming more prevalent in IT systems and is used by
enterprises and organizations in a variety of business intelligence
and data analytics applications. Data mining and machine learning,
such as natural language processing and text analytics, provide
different techniques to find patterns in, or otherwise interpret,
this unstructured information. Unstructured data typically resides
in a graph database.
A graph database is a database that uses graph structures for
semantic queries with nodes, edges, and properties to represent and
store data. Nodes in the graph database represent entities, such
as, for example, people, businesses, accounts, products, objects,
or any other item you might want to keep track of. Properties are
pertinent information that relate to nodes. Edges represent the
relationships that connect nodes to nodes or nodes to properties.
The edges may be directed from one node to another or undirected
with no specific from-to relationship between a pair of nodes.
The graph database stores the unstructured data according to a
graph schema. The graph schema represents a logical configuration
of the graph database. Thus, the graph schema indicates how
entities that make up the graph database relate to one another. In
other words, at a most basic level, the graph schema indicates
which data make up the graph database.
Retrieving data from the graph database requires a query language,
such as, for example, SPARQL (SPARQL Protocol and RDF Query
Language). SPARQL is a Resource Description Framework (RDF) query
language, which is a semantic query language for graph databases,
able to retrieve and manipulate data stored in an RDF format.
In today's IT environment, professionals are faced with management
of vast amounts of data (e.g., Big Data). Choices have to made with
respect to modeling and storing this data. The use of a relational
database may indicate that the relationships and entities within
the data are of a relatively static disposition and are well
understood by the domain experts. The use of a graph database may
indicate that the domain of knowledge is too large to understand
completely, and too dynamic (e.g., changing constantly or within
short intervals of time) to model in a relational paradigm, where
schema changes can have a severe impact on dependent
applications.
SUMMARY
According to one illustrative embodiment, a computer-implemented
method for detecting missing entities in a graph schema is
provided. Unlabeled and unstructured data of a particular domain
are divided into a plurality of Voronoi cells using a clustering
algorithm that is initiated with cluster centroids proportional to
a number of entity types corresponding to the particular domain
existing in the graph schema. One additional cluster more than the
number of entity types corresponding to the particular domain
existing in the graph schema is initialized using a cluster
initializing formula of the clustering algorithm. It is determined
whether the one additional cluster is populated. In response to
determining that the one additional cluster is populated, an entity
type is determined to be missing from the number of entity types
existing in the graph schema and the missing entity type is added
to the graph schema. In response to determining that the one
additional cluster is not populated, all entity types of the
particular domain are determined to exist in the graph schema.
In addition, the number of entity types existing in the graph
schema is incremented by one accounting for the added missing
entity type. The clustering algorithm is then re-executed
initiating a number of cluster centroids proportional to the
incremented number of entity types. A number of clusters
corresponding to the incremented number of entity types are
populated and a determination is made as to whether a cluster in
the number of clusters is not populated. In response to determining
that the number of clusters is populated, another entity type is
determined to be missing from the graph schema. In response to
determining that a cluster in the number of clusters is not
populated, all entity types of the particular domain are determined
to exist in the graph schema. Further, the graph schema having all
entity types corresponding to the particular domain modeled in the
graph schema is outputted to a graph database server hosting a
graph database that complies with the graph schema to increase
query performance and efficiency of the graph database server.
Also, a notification regarding the missing entity type is sent.
According to other illustrative embodiments, a computer system and
computer program product for detecting missing entities in a graph
schema are provided.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a pictorial representation of a network of data
processing systems in which illustrative embodiments may be
implemented;
FIG. 2 is a diagram of a data processing system in which
illustrative embodiments may be implemented; and
FIG. 3 is a flowchart illustrating a process for assessing
correctness of entities corresponding to a particular domain
existing in a graph schema by detecting missing entity types in the
graph schema and automatically adding the detected missing entity
types in accordance with an illustrative embodiment.
DETAILED DESCRIPTION
The present invention may be a system, a method, and/or a computer
program product at any possible technical detail level of
integration. The computer program product may include a computer
readable storage medium (or media) having computer readable program
instructions thereon for causing a processor to carry out aspects
of the present invention.
The computer readable storage medium can be a tangible device that
can retain and store instructions for use by an instruction
execution device. The computer readable storage medium may be, for
example, but is not limited to, an electronic storage device, a
magnetic storage device, an optical storage device, an
electromagnetic storage device, a semiconductor storage device, or
any suitable combination of the foregoing. A non-exhaustive list of
more specific examples of the computer readable storage medium
includes the following: a portable computer diskette, a hard disk,
a random access memory (RAM), a read-only memory (ROM), an erasable
programmable read-only memory (EPROM or Flash memory), a static
random access memory (SRAM), a portable compact disc read-only
memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a
floppy disk, a mechanically encoded device such as punch-cards or
raised structures in a groove having instructions recorded thereon,
and any suitable combination of the foregoing. A computer readable
storage medium, as used herein, is not to be construed as being
transitory signals per se, such as radio waves or other freely
propagating electromagnetic waves, electromagnetic waves
propagating through a waveguide or other transmission media (e.g.,
light pulses passing through a fiber-optic cable), or electrical
signals transmitted through a wire.
Computer readable program instructions described herein can be
downloaded to respective computing/processing devices from a
computer readable storage medium or to an external computer or
external storage device via a network, for example, the Internet, a
local area network, a wide area network and/or a wireless network.
The network may comprise copper transmission cables, optical
transmission fibers, wireless transmission, routers, firewalls,
switches, gateway computers and/or edge servers. A network adapter
card or network interface in each computing/processing device
receives computer readable program instructions from the network
and forwards the computer readable program instructions for storage
in a computer readable storage medium within the respective
computing/processing device.
Computer readable program instructions for carrying out operations
of the present invention may be assembler instructions,
instruction-set-architecture (ISA) instructions, machine
instructions, machine dependent instructions, microcode, firmware
instructions, state-setting data, configuration data for integrated
circuitry, or either source code or object code written in any
combination of one or more programming languages, including an
object oriented programming language such as Smalltalk, C++, or the
like, and procedural programming languages, such as the "C"
programming language or similar programming languages. The computer
readable program instructions may execute entirely on the user's
computer, partly on the user's computer, as a stand-alone software
package, partly on the user's computer and partly on a remote
computer or entirely on the remote computer or server. In the
latter scenario, the remote computer may be connected to the user's
computer through any type of network, including a local area
network (LAN) or a wide area network (WAN), or the connection may
be made to an external computer (for example, through the Internet
using an Internet Service Provider). In some embodiments,
electronic circuitry including, for example, programmable logic
circuitry, field-programmable gate arrays (FPGA), or programmable
logic arrays (PLA) may execute the computer readable program
instructions by utilizing state information of the computer
readable program instructions to personalize the electronic
circuitry, in order to perform aspects of the present
invention.
Aspects of the present invention are described herein with
reference to flowchart illustrations and/or block diagrams of
methods, apparatus (systems), and computer program products
according to embodiments of the invention. It will be understood
that each block of the flowchart illustrations and/or block
diagrams, and combinations of blocks in the flowchart illustrations
and/or block diagrams, can be implemented by computer readable
program instructions.
These computer readable program instructions may be provided to a
processor of a general purpose computer, special purpose computer,
or other programmable data processing apparatus to produce a
machine, such that the instructions, which execute via the
processor of the computer or other programmable data processing
apparatus, create means for implementing the functions/acts
specified in the flowchart and/or block diagram block or blocks.
These computer readable program instructions may also be stored in
a computer readable storage medium that can direct a computer, a
programmable data processing apparatus, and/or other devices to
function in a particular manner, such that the computer readable
storage medium having instructions stored therein comprises an
article of manufacture including instructions which implement
aspects of the function/act specified in the flowchart and/or block
diagram block or blocks.
The computer readable program instructions may also be loaded onto
a computer, other programmable data processing apparatus, or other
device to cause a series of operational steps to be performed on
the computer, other programmable apparatus or other device to
produce a computer implemented process, such that the instructions
which execute on the computer, other programmable apparatus, or
other device implement the functions/acts specified in the
flowchart and/or block diagram block or blocks.
The flowchart and block diagrams in the Figures illustrate the
architecture, functionality, and operation of possible
implementations of systems, methods, and computer program products
according to various embodiments of the present invention. In this
regard, each block in the flowchart or block diagrams may represent
a module, segment, or portion of instructions, which comprises one
or more executable instructions for implementing the specified
logical function(s). In some alternative implementations, the
functions noted in the blocks may occur out of the order noted in
the Figures. For example, two blocks shown in succession may, in
fact, be executed substantially concurrently, or the blocks may
sometimes be executed in the reverse order, depending upon the
functionality involved. It will also be noted that each block of
the block diagrams and/or flowchart illustration, and combinations
of blocks in the block diagrams and/or flowchart illustration, can
be implemented by special purpose hardware-based systems that
perform the specified functions or acts or carry out combinations
of special purpose hardware and computer instructions.
With reference now to the figures, and in particular, with
reference to FIG. 1 and FIG. 2, diagrams of data processing
environments are provided in which illustrative embodiments may be
implemented. It should be appreciated that FIG. 1 and FIG. 2 are
only meant as examples and are not intended to assert or imply any
limitation with regard to the environments in which different
embodiments may be implemented. Many modifications to the depicted
environments may be made.
FIG. 1 depicts a pictorial representation of a network of data
processing systems in which illustrative embodiments may be
implemented. Network data processing system 100 is a network of
computers, data processing systems, and other devices in which the
illustrative embodiments may be implemented. Network data
processing system 100 contains network 102, which is the medium
used to provide communications links between the computers, data
processing systems, and other devices connected together within
network data processing system 100. Network 102 may include
connections, such as, for example, wire communication links,
wireless communication links, and fiber optic cables.
In the depicted example, server 104 and server 106 connect to
network 102, along with domain storage 108. Server 104 and server
106 may be, for example, server computers with high-speed
connections to network 102. In addition, it should be noted that
server 104 and server 106 may each represent clusters of servers in
one or more data centers. Alternatively, server 104 and server 106
may each represent multiple computing nodes in a cloud
environment.
Also in the depicted example, server 104 is a computer that
analyzes a graph schema corresponding to a graph database hosted by
graph database server 106 in order to detect missing entity types,
which correspond to a particular knowledge domain, within the graph
schema. Server 104 ingests and analyzes data 110 of domain storage
108. Data 110 may be, for example, unlabeled and unstructured data
corresponding to a particular domain, such as an oil industry
domain, an insurance industry domain, a product manufacturing
domain, an education domain, a healthcare domain, an automotive
domain, a government domain, a food service domain, a legal domain,
or the like. However, it should be noted that data 110 also may
include structured data as well. Server 104 analyzes ingested data
110 using, for example, machine learning. Machine learning is a
subset of artificial intelligence that uses statistical techniques
to give server 104 the ability to learn from ingested data 110
without being explicitly programmed to do so.
Based on the analysis, server 104, using a clustering algorithm,
divides ingested data 100 into a plurality of clusters and adds one
additional cluster. Each of the plurality of clusters corresponds
to an existing entity in the graph schema. If server 104 detects
that the additional cluster is populated with data, then server 104
determines that an entity type is missing from the graph schema
corresponding to the graph database of graph database server 106.
Server 104 then adds the missing entity type to the graph schema.
By ensuring that the graph schema is updated and complete and that
the graph database of graph database server 106 complies with the
updated graph schema, server 104 is able to increase the query
performance and efficiency of graph database server 106.
Domain storage 108 represents a network storage device capable of
storing any type of domain information in an unstructured and a
structured format. In addition, domain storage 108 may represent a
plurality of different local and/or remote network storage devices
that store information for a set of one or more knowledge domains.
Further, domain storage 108 may store other types of information,
such as authentication or credential data that may include user
names, passwords, and biometric data associated with domain subject
matter experts or system administrators, for example.
Client 112, client 114, and client 116 also connect to network 102.
Clients 112, 114, and 116 are clients of graph database server 106.
In this example, clients 112, 114, and 116 are shown as desktop or
personal computers with wire communication links to network 102.
However, it should be noted that clients 112, 114, and 116 are
examples only and may represent other types of data processing
systems, such as, for example, network computers, laptop computers,
handheld computers, smart phones, smart watches, smart televisions,
kiosks, and the like. Users of clients 112, 114, and 116 may
utilize clients 112, 114, and 116 to submit data queries requesting
domain information from graph database server 106.
In addition, it should be noted that network data processing system
100 may include any number of additional servers, clients, storage
devices, and other devices not shown. Program code located in
network data processing system 100 may be stored on a computer
readable storage medium and downloaded to a computer or other data
processing device for use. For example, program code may be stored
on a computer readable storage medium on server 104 and downloaded
to graph database server 106 over network 102 for use on graph
database server 106.
In the depicted example, network data processing system 100 may be
implemented as a number of different types of communication
networks, such as, for example, an internet, an intranet, a local
area network (LAN), a wide area network (WAN), or any combination
thereof. FIG. 1 is intended as an example only, and not as an
architectural limitation for the different illustrative
embodiments.
With reference now to FIG. 2, a diagram of a data processing system
is depicted in accordance with an illustrative embodiment. Data
processing system 200 is an example of a computer, such as server
104 in FIG. 1, in which computer readable program code or
instructions implementing processes of illustrative embodiments may
be located. In this illustrative example, data processing system
200 includes communications fabric 202, which provides
communications between processor unit 204, memory 206, persistent
storage 208, communications unit 210, input/output (I/O) unit 212,
and display 214.
Processor unit 204 serves to execute instructions for software
applications and programs that may be loaded into memory 206.
Processor unit 204 may be a set of one or more hardware processor
devices or may be a multi-processor core, depending on the
particular implementation.
Memory 206 and persistent storage 208 are examples of storage
devices 216. A computer readable storage device is any piece of
hardware that is capable of storing information, such as, for
example, without limitation, data, computer readable program code
in functional form, and/or other suitable information either on a
transient basis and/or a persistent basis. Further, a computer
readable storage device excludes a propagation medium. Memory 206,
in these examples, may be, for example, a random-access memory
(RAM), or any other suitable volatile or non-volatile storage
device. Persistent storage 208 may take various forms, depending on
the particular implementation. For example, persistent storage 208
may contain one or more devices. For example, persistent storage
208 may be a hard drive, a flash memory, a rewritable optical disk,
a rewritable magnetic tape, or some combination of the above. The
media used by persistent storage 208 may be removable. For example,
a removable hard drive may be used for persistent storage 208.
In this example, persistent storage 208 stores graph schema missing
entity detector 218. However, it should be noted that even though
graph schema missing entity detector 218 is illustrated as residing
in persistent storage 208, in an alternative illustrative
embodiment graph schema missing entity detector 218 may be a
separate component of data processing system 200. For example,
graph schema missing entity detector 218 may be a hardware
component coupled to communication fabric 202 or a combination of
hardware and software components. In another alternative
illustrative embodiment, a first portion of graph schema missing
entity detector 218 may be located in data processing system 200
and a second portion of graph schema missing entity detector 218
may be located in a second data processing system, such as, for
example, graph database server 106 in FIG. 1.
Graph schema missing entity detector 218 controls the process of
assessing correctness of existing entities, which correspond to a
particular domain, within a graph schema by detecting missing
entity types in the graph schema and automatically adding the
missing entity types to the graph schema to increase graph database
query performance and efficiency. Graph database 220 represents a
listing of a set of one or more domain knowledge graph databases
residing in the graph database server. Graph database 220 contains
information corresponding to a particular domain of knowledge.
Graph database 220 complies with graph schema 222. Graph schema 222
includes root entities 224 and entity types 226. Root entities 224
represent a set of one or more entities modeled in graph schema
222. Root entities 224 have no parent entities. In other words,
root entities 224 represent root nodes in graph database 220.
Entity types 226 represent a plurality of sub-entities that are in
a child relationship with one or more of root entities 224. In
other words, entity types 226 represent child nodes of parent nodes
associated with root entities 224 in graph database 220.
Knowledge domain 228 represents a particular domain of knowledge
stored in graph database 220. Knowledge domain 228 may represent
any domain of knowledge. Data sources 230 represent a plurality of
local and remote sources of information corresponding to the
particular domain of knowledge represented by knowledge domain 228.
Data sources 230 include unlabeled and unstructured data 232.
However, it should be noted that data sources 230 also may include
labeled and structured data in addition to unlabeled and
unstructured data 232.
Graph schema missing entity detector 218 utilizes machine learning
component 234 to analyze unlabeled and unstructured data 232. Graph
schema missing entity detector 218 utilizes clustering algorithm
236 to partition unlabeled and unstructured data 232 into clusters
238 based on the analysis of unlabeled and unstructured data 232 by
machine learning component 234. Clusters 238 are proportional to
root entities 224. In addition, clustering algorithm 236 includes
clustering initializing formula 240, which initializes one
additional cluster, such as additional cluster 242.
If graph schema missing entity detector 218 detects that additional
cluster 242 is populated with data points, then graph schema
missing entity detector 218 determines that an entity type is
missing from graph schema 222 and adds the missing entity type to
graph schema 222. Afterward, graph schema missing entity detector
218 outputs updated graph schema 222 to the graph database server
hosting graph database 220 for implementation.
Communications unit 210, in this example, provides for
communication with other computers, data processing systems, and
devices via a network, such as network 102 in FIG. 1.
Communications unit 210 may provide communications through the use
of both physical and wireless communications links. The physical
communications link may utilize, for example, a wire, cable,
universal serial bus, or any other physical technology to establish
a physical communications link for data processing system 200. The
wireless communications link may utilize, for example, shortwave,
high frequency, ultra high frequency, microwave, wireless fidelity
(Wi-Fi), Bluetooth.RTM. technology, global system for mobile
communications (GSM), code division multiple access (CDMA),
second-generation (2G), third-generation (3G), fourth-generation
(4G), 4G Long Term Evolution (LTE), LTE Advanced, or any other
wireless communication technology or standard to establish a
wireless communications link for data processing system 200.
Input/output unit 212 allows for the input and output of data with
other devices that may be connected to data processing system 200.
For example, input/output unit 212 may provide a connection for
user input through a keypad, a keyboard, a mouse, a microphone,
and/or some other suitable input device. Display 214 provides a
mechanism to display information to a user and may include touch
screen capabilities to allow the user to make on-screen selections
through user interfaces or input data, for example.
Instructions for the operating system, applications, and/or
programs may be located in storage devices 216, which are in
communication with processor unit 204 through communications fabric
202. In this illustrative example, the instructions are in a
functional form on persistent storage 208. These instructions may
be loaded into memory 206 for running by processor unit 204. The
processes of the different embodiments may be performed by
processor unit 204 using computer-implemented instructions, which
may be located in a memory, such as memory 206. These program
instructions are referred to as program code, computer usable
program code, or computer readable program code that may be read
and run by a processor in processor unit 204. The program
instructions, in the different embodiments, may be embodied on
different physical computer readable storage devices, such as
memory 206 or persistent storage 208.
Program code 244 is located in a functional form on computer
readable media 246 that is selectively removable and may be loaded
onto or transferred to data processing system 200 for running by
processor unit 204. Program code 244 and computer readable media
246 form computer program product 248. In one example, computer
readable media 246 may be computer readable storage media 250 or
computer readable signal media 252. Computer readable storage media
250 may include, for example, an optical or magnetic disc that is
inserted or placed into a drive or other device that is part of
persistent storage 208 for transfer onto a storage device, such as
a hard drive, that is part of persistent storage 208. Computer
readable storage media 250 also may take the form of a persistent
storage, such as a hard drive, a thumb drive, or a flash memory
that is connected to data processing system 200. In some instances,
computer readable storage media 250 may not be removable from data
processing system 200.
Alternatively, program code 244 may be transferred to data
processing system 200 using computer readable signal media 252.
Computer readable signal media 252 may be, for example, a
propagated data signal containing program code 244. For example,
computer readable signal media 252 may be an electro-magnetic
signal, an optical signal, and/or any other suitable type of
signal. These signals may be transmitted over communication links,
such as wireless communication links, an optical fiber cable, a
coaxial cable, a wire, and/or any other suitable type of
communications link. In other words, the communications link and/or
the connection may be physical or wireless in the illustrative
examples. The computer readable media also may take the form of
non-tangible media, such as communication links or wireless
transmissions containing the program code.
In some illustrative embodiments, program code 244 may be
downloaded over a network to persistent storage 208 from another
device or data processing system through computer readable signal
media 252 for use within data processing system 200. For instance,
program code stored in a computer readable storage media in a data
processing system may be downloaded over a network from the data
processing system to data processing system 200. The data
processing system providing program code 244 may be a server
computer, a client computer, or some other device capable of
storing and transmitting program code 244.
The different components illustrated for data processing system 200
are not meant to provide architectural limitations to the manner in
which different embodiments may be implemented. The different
illustrative embodiments may be implemented in a data processing
system including components in addition to, or in place of, those
illustrated for data processing system 200. Other components shown
in FIG. 2 can be varied from the illustrative examples shown. The
different embodiments may be implemented using any hardware device
or system capable of executing program code. As one example, data
processing system 200 may include organic components integrated
with inorganic components and/or may be comprised entirely of
organic components excluding a human being. For example, a storage
device may be comprised of an organic semiconductor.
As another example, a computer readable storage device in data
processing system 200 is any hardware apparatus that may store
data. Memory 206, persistent storage 208, and computer readable
storage media 246 are examples of physical storage devices in a
tangible form.
In another example, a bus system may be used to implement
communications fabric 202 and may be comprised of one or more
buses, such as a system bus or an input/output bus. Of course, the
bus system may be implemented using any suitable type of
architecture that provides for a transfer of data between different
components or devices attached to the bus system. Additionally, a
communications unit may include one or more devices used to
transmit and receive data, such as a modem or a network adapter.
Further, a memory may be, for example, memory 206 or a cache such
as found in an interface and memory controller hub that may be
present in communications fabric 202.
A graph formalization (i.e., graph schema), such as, for example, a
web ontology, is a description of entities that exist is a
particular domain of knowledge and the relationships between the
entities within that particular domain. A graph database is a
collection of instance data that conforms to a set of one or more
graph schemas. Entities within a graph schema may be either
manually modeled by a subject matter expert of a particular domain
or automatically modeled by a computer.
In cases where a graph database is utilized to store data,
techniques for building out a graph schema for that graph database
may feature an automated component. For example, a subject matter
expert may manually build the graph schema and then the graph
schema is supplemented through computer-automated techniques. In
other cases, or in situations where noise is more readily
tolerated, the entire schema may be developed through an automated
approach. In either approach, it is important to assess the
correctness of the entities that have been modeled in the schema
and make suggestions for new entity types, where some entity types
may have been missed. It also may be important to indicate where an
entity may be over-generalized. In cases of entity over
generalization, an over-generalized entity can be further specified
and decomposed into multiple sub-entities (i.e., entity types) or
sub-classes. In other situations, an entity may be over-specified
and may need to be more generalized. Either error (i.e.,
over-generalized or under-specified) is possible, whether
developing a graph schema manually or automatically, and should be
corrected for a more efficient data model and an accurate
reflection of the knowledge in that domain.
Illustrative embodiments automatically detect entity gaps within an
existing graph schema for the purpose of adding new entity types to
the existing graph schema and automatically assessing accuracy of
existing entities in the graph schema. As an example scenario, in
an oil and gas industry domain, a Reservoir entity may be
characterized by frequent links to Depth and Porosity entities. The
Depth entity may be characterized by having entity types, such as a
True Vertical Depth entity type, a Mean Sea Level entity type, and
a Median Sea Level entity type. It is the job of a graph database
to hold instance data corresponding to that particular data model
for that particular domain. The data conforming to that particular
model may be logically represented as, for example: alpha rdf:type
Reservoir; alpha has (3000 m rdf:type Mean Sea Level).
In this particular oil and gas industry domain example, a subject
matter expert would be able to point to this data model and
indicate that information is missing. In this example, the entity
Reservoir is under-specified. For example, many types of
Reservoirs, such as Carbonate Reservoirs, Sandstone Reservoirs, and
the like, exist that should be modeled to effectively understand
the relationship to the Depth entity. Similarly, the Depth entity
is over-specified. While True Vertical Depth and Mean Sea Level are
common depth designations, the entity specification in the graph
schema of Median Sea Level may have been the result of an
over-zealous subject matter expert in modeling the data. For
example, the Median Sea Level entity type may have less relevance
to the Depth entity.
However, current computer algorithms are not necessarily going to
arrive at the same conclusion as a subject matter expert with
respect to this data model. For example, this data model is both
syntactically and semantically valid. In computational terms, this
data model can be compiled and data can be associated with and
retrieved from this data model.
Furthermore, a manual review of this data model by a subject matter
expert may indicate that the Porosity entity is missing. The
association of Depth entities to Reservoir entities is not
incorrect, but this association will take on greater meaning during
any analysis of this data model if the Porosity entity also is
included. Finally, the subject matter expert also will likely
assess accuracy of all entities in the graph schema. In this
example, all entities are relevant, but relevance can be a
subjective concern particularly in larger data models.
Illustrative embodiments detect and add new entity types to graph
schemas through analysis of unlabeled and unstructured data of a
particular domain using machine learning. By using a clustering
algorithm, which is initiated with cluster centroids proportional
to root entities of a particular domain in a graph schema,
illustrative embodiments divide the unlabeled and unstructured data
into Voronoi cells. The clustering algorithm may utilize a cluster
initialization formula, such as, for example: a
.sigma..varies.(.omega.+1) (i.e., the number of clustering
centroids is proportional to the number of root entities plus one).
Using the example above, illustrative embodiments assume that both
Reservoir and Depth entities have already been modeled and that
instance data for these entities exist. Illustrative embodiments
initialize one additional cluster using the cluster initialization
formula. In other words, the cluster initialization formula will
always initialize the cluster centroids with one cluster more than
entities that exist in the graph schema. If this one additional
cluster is populated, it indicates that an entity type is missing
from the graph schema. Using the above example, illustrative
embodiments assume that the clustered data corresponding to the one
additional cluster is indicative of the Porosity entity, which was
not previously modeled. The data that was ingested by illustrative
embodiments from the oil and gas industry domain contain multiple
relationships from the Reservoir entity to the Porosity entity and
from the Depth entity to the Porosity entity. Based on the results
of the clustering algorithm, the Porosity entity is clearly a
separate entity and one that has not been modeled in the existing
graph schema.
Once illustrative embodiments detect the missing entity,
illustrative embodiments execute the clustering algorithm once
again. In other words, illustrative embodiments repeat the cluster
initialization formula. This time, illustrative embodiments
increment the number of existing entities within the graph schema
by one (i.e., the Porosity entity has been added to the graph
schema). Therefore, using the above example, illustrative
embodiments will initialize the cluster centroids using a value of
four. The initialization of four cluster centroids will result in
the population of only three clusters, which indicates that no
other missing entities exist in the graph schema.
However, it should be noted that the detection of missing entities
depends upon the extent of the source domain data available. In
other words, machine learning algorithms are dependent upon the
amount of source data provided to them. While this may be taken as
a limitation in some cases, it should be noted that in other cases
it becomes a preferred approach. The entities modeled in a graph
schema will never be greater than the source domain data behind the
modeled entities.
With reference now to FIG. 3, a flowchart illustrating a process
for assessing correctness of entities corresponding to a particular
domain existing in a graph schema by detecting missing entity types
in the graph schema and automatically adding the detected missing
entity types is shown in accordance with an illustrative
embodiment. The process shown in FIG. 3 may be implemented in a
computer, such as server 104 in FIG. 1 or data processing system
200 in FIG. 1.
The process begins when the computer ingests unlabeled and
unstructured data corresponding to a particular domain from a
plurality of data sources via a network (step 302). The computer
analyzes the unlabeled and unstructured data corresponding to the
particular domain using machine learning (step 304). Based on the
analysis, the computer divides the unlabeled and unstructured data
into a plurality of Voronoi cells using a clustering algorithm that
is initiated with cluster centroids proportional to a number of
entity types corresponding to the particular domain existing in a
graph schema (step 306). The clustering algorithm may be, for
example, a k-means clustering algorithm. However, alternative
illustrative embodiments are not limited to such. In other words,
different illustrative embodiments may utilize one or more
different types of clustering algorithms. For example, alternative
illustrative embodiments may utilize mean-shift clustering,
density-based spatial clustering, expectation-maximization
clustering using Gaussian Mixture Models, hierarchical clustering,
or the like.
Further, the computer initializes one additional cluster more than
the number of entity types corresponding to the particular domain
existing in the graph schema using a cluster initializing formula
of the clustering algorithm (step 308). The clustering initializing
formula may be, for example, .sigma..varies.(.omega.+1), where a is
equal to a number of clustering centroids and co is equal to a
number of root entities in a graph schema. In other words, the
number of clustering centroids is proportional to the number of
root entities plus one. A root entity is an entity that has no
parent entity. An entity type is a sub-entity that is in a child
relationship to a root entity.
Furthermore, the computer populates clusters corresponding to
entity types of the particular domain based on dividing the
unlabeled and unstructured data into the plurality of Voronoi cells
(step 310). Afterward, the computer makes a determination as to
whether the one additional cluster is populated (step 312). If the
computer determines that the one additional cluster is not
populated, no output of step 312, then the process proceeds to step
326. If the computer determines that the one additional cluster is
populated, yes output of step 312, then the computer determines
that an entity type is missing from the number of entity types
existing in the graph schema (step 314).
The computer adds the missing entity type to the graph schema (step
316). By adding the missing entity type to the graph schema, the
computer transforms the graph schema into a new and modified graph
schema. In an alternative illustrative embodiment, instead of, or
in addition to, adding the missing entity type to the graph schema,
the computer generates and sends a notification to a user, such as
a database administrator, regarding the missing entity type for
review and possible action.
After adding the missing entity type to the graph schema in step
316, the computer increments the number of entity types existing in
the graph schema by one accounting for the added missing entity
type (step 318). Then, the computer re-executes the clustering
algorithm initiating a number of cluster centroids proportional to
the incremented number of entity types (step 320).
The computer populates a number of clusters corresponding to the
incremented number of entity types (step 322). Afterward, the
computer makes a determination as to whether a cluster in the
number of clusters is not populated (step 324). If the computer
determines that the number of clusters is populated, no output of
step 324, then the process returns to step 314 where the computer
determines that another entity type is missing. If the computer
determines that a cluster in the number of clusters is not
populated, yes output of step 324, then the computer determines
that all entity types of the particular domain exist in the graph
schema (step 326). Subsequently, the computer outputs the graph
schema having all of the entity types corresponding to the
particular domain modeled in the graph schema (step 328). The
computer outputs the graph schema to a graph database server, such
as graph database server 106 in FIG. 1, to implement the graph
database for that particular domain and increase query performance
of that particular graph database server. In addition, the computer
may output the graph schema to a graph database administrator for
review. Thereafter, the process terminates.
Thus, illustrative embodiments of the present invention provide a
computer-implemented method, computer system, and computer program
product for assessing correctness of entities corresponding to a
particular domain existing in a graph schema by detecting missing
entity types in the graph schema and automatically adding the
detected missing entity types to the graph schema. The descriptions
of the various embodiments of the present invention have been
presented for purposes of illustration, but are not intended to be
exhaustive or limited to the embodiments disclosed. Many
modifications and variations will be apparent to those of ordinary
skill in the art without departing from the scope and spirit of the
described embodiments. The terminology used herein was chosen to
best explain the principles of the embodiments, the practical
application or technical improvement over technologies found in the
marketplace, or to enable others of ordinary skill in the art to
understand the embodiments disclosed herein.
* * * * *
References